Search

The effect of segregation behavior of non-metallic dopants (H/He/O) and metallic dopants (Be/Al/Mg/Y) on the performance of grain boundary (GB) in SiC has been systematically investigated by first-principles calculations. Firstly, the GB energy and excess volume of different GBs have been studied to evaluate the stability of GB and the capacity to accommodate dopant atoms. The solution energies of dopant atoms greatly reduce in the GB region compared with those in the bulk, which makes the dopant atoms inside the grain tend to segregate and aggregate near the GB. The driving force of GB on dopant segregation generally decreases with the increase of distance from GB plane, and the preferential site of dopant is closely correlated with the atomic size of dopant. In addition, H and Y atom possesses the lowest segregation energy at the interstitial and substitutional site near the GB, respectively. Next, the segregation of single dopant induced the changes in the strength and stability of GB have been explored. It is found that non-metallic dopants have the significant embrittlement effects on GB strength. However, the segregation of most metallic dopants could strengthen the GB and Mg atom has the most significant strengthening effect on the GB. The stability of GB can be greatly improved by segregation of Al and Y dopants. Besides, the aggregation of H atoms has the obvious embrittlement effect on the GB. Furthermore, the co-segregation behavior of different dopants has also been explored. Be and Mg dopants have the most significant inhibition effect on the segregation of detrimental impurities H/He/O due to the repulsive interaction between dopant atoms. The present results provide a new insight into the effect of dopant segregation on GB properties and are expected to be a useful guidance for screening the chemical composition and manipulating the performance of SiC-based ceramics. [ABSTRACT FROM AUTHOR]

[Display omitted] Currently, neural networks have become an attractive tool for achieving highly quantitative accuracy without time-consuming fitting. In this study, we constructed a neural network potential (NNP) for silicon carbide (SiC) by employing neural networks to study the elastic response. The results showed that, compared to existing empirical potential, the NNP for SiC demonstrated greater accuracy in calculating various material properties. Additionally, this potential successfully reproduced the tension-strain behavior of monocrystalline SiC. In describing polycrystalline SiC, there is a limit to the transferability of the NNP, particularly in terms of its dynamical response. This defect could be improved by introducing additional corresponding training samples. This study not only provides a reliable SiC potential, but also contributes to assessing the applicability and limitation of NNPs. In addition, we place emphasis on the significance of the composition of the training set and the optimization of training weights in NNP training. [ABSTRACT FROM AUTHOR]